“Mindfulness of Breathing in the Dhyāna Sūtras.” by Florin Deleanu Transactions of the International Conference of Orientalists in Japan (TICOJ) 37, 1992, 42-57

Research Reports

Mindfulness of Breathing in the Dhyiina Sutras

Florin DELEANU

Together with the meditation on the impure (asubhasmrti), mindful­ ness of breathing (Pali, iiniipiinasati; Sanskrit, iiniipiinasmrti; variously rendered into Chinese as 'R~)t~~, 'R~)t~, JliiJjjG~jjG1iJl" ~I±l)\,~" ¥,f,~,:@;, ~,~,1iJl" etc.) represents one of the main forms of spiritual cul­ tivation in the Early and Schismatic Buddhism.!) Abhidharma liter­ ature often celebrates these two meditative practices as " the two gates of ambrosia (amrta) i=Jar~,"2) "the two main gates of entering spir­ itual cultivation )\{~~=r~,"3) and" the two paths leading to Nirva.g.a Jl@i1E~='f.I."4) Mahayana Buddhism, most notably the Representa­ tion-only School,5) retains mindfulness of breathing amongst its spir­ itual techniques, but its role is much diminished as the dominant posi-

1) In what follows I shall avoid the usage of the rather pejorative term of Hinayana or Lesser Vehicle and replace it with Schismatic Buddhism or Con­ servative Buddhism. 2) Abhidharma mahavibhlifa-sastra JliiJmJ¥~*l'M~?~ilifil (Taisho shinshu dai­ zo-kyo [=TJ) vol. 27, p. 348b16-17), Pancavastu vibhiiia-sastra ::n:.m~rY ilifil (T28, 989c). 3) Abhidharmakosa-bhlifya JliiJmJ¥~{j'!,4rili1il (T29, 117b6), Abhidharma nya­ yiinusiira-siistra JliiJmJ¥~]IIi'i1Eili1il (T29, 671a1). 4) Abhidharmamrta-sastra JliiJm.if4,*~ffiI (T28, 975bl1-12). 5) iiVe find, for instance, the mindfulness of breathing extensively discussed in the Yogacaryabhumi :fJl«fiJuSi!i±fuilifil (T30, 430c-433b). However, we must note that this fragment is included in the Sravakabhumi and the Boddhisattvabhumi contains only a brief mention of it: "vitarkapratipak~en'anapana-smrtirp." (Bodhisattvabhumi, ed. by Unrai Wogihara, Tokyo, 1930, p. 204); 7.~:':l~kMtii~ ~Ji!!,~{!t11:W,~,~ (T30, 527a2).

( 42 ) tion is now occupied by devotional practices, visualization exercises, meditation aimed at the realization of emptiness, etc. The purpose of the present paper is to examine the treatment of the mindfulness of breathing, especially the relation between" the sixteen bases" and" the six aspects," in the so-called dhyiina siitras or " med­ itation scriptures." The term dhyiina siitra6) represents a reconstruc­ tion of the Chinese chan jing *,-;f,¥, which appears in the titles of a certain group of texts as well as in the writings of Chinese Buddhists of the early 5th century AD as Sengrui {~®1,7) Huiyuan ~~,8) and

Huiguan ~M. 9) This group represents treatises or manuals of meditation belonging to or, at least, partly drawing their inspiration from Sarvastivada tradi­ tion. We must not forget, however, that many of the dhyiina siitras are, as it were, a common product of the Buddhist traditions from North-West India, Central Asia, and China. Although the basic ma­ terial of these texts doubtlessly comes from Indian Buddhism, we must not exclude the possibility that they were compiled in Central Asia or that Buddhist monks from Central Asia, who in most of the cases were those who brought the dhyiina siitras to China, had a more or less sub­ stantial influence on the final form of the scriptures in the process of translating them into Chinese. Kumarajrva would be perhaps the most eloquent example of this case.

6) I could find no evidence in Indian sources and Tibetan translations that the chan jing stands for an original dhyanasutrani or something similar. The dhyana siitras represents only a reconstruction whose only advantage is that it suggests better that we are actually dealing with a category of texts which, regardless of their compilation process, contain doctrines and practices that can be traced back to Indian Buddhism. 7) V. T55, 65a-b. Sengrui also uses the term chan yao tli!~ or "medita­ tion summaries." 8) V. T55, 65b-66a. 9) V. T55, 66b-67c. Huiguan also calls such texts chan dian:/lliJlll, or "med­ itation scriptures."

( 43 ) Doctrinally, the dhyiina siltras range from a basically orthodox Sar­ vastivada standpoint to a substantial compromise with Mahayana teach­ ings and practices. In some cases, this compromise is so advanced that it is very difficult to make a distinction between such a text and a samadhi siUra (sanmei jing .::::.!!*g), which represents a purely Maha­ yanist scripture. 10) Texts like An Shigao's *i:!tfeli (fl. 148-170 AD) translations: the Da anban shouyijing **~5t"'f;i:g (T1S, 163c-173a), the Skandhadhiitviiyatana-siitra ~Ff.Ag (T1S, 173b-180b), Sarpgha­ rak~a's Yogiiciirabhiimi-siltra *J!t:Li!!,g (T1S, 230c-236b) (partial trans­ lation); the first 27 chapters & of the present version of Dharmarak~a's ~~~ (239-316) translation of Sarpgharak~a's YogiiciirabhzImi-siitra {~ 1TJ!t±llig (T1S, 181c-223a); Buddhabhadra's 1~1£tJiW:~t!f,i (359-429) translation of Buddhasena's Yogiiciirabhiimi-szItra, also called the Dhar­ matiira dhyiina-siltra or Dharmatriita dhyiina-siltra 3i}li,:~!f,i1lj!.g (T1S, 300c-32Sc); etc. belong to an orthodox Sarvastivada position. On the other hand, we have dhyiina siitras that combine to various extents the meditative system of the Conservative Buddhism with Mahayana med­ itation and teachings. Their way of mixing the two systems is not uniform. Kumarajiva's Mb}li,:!f,i1t (344-412) translation or compilation of the Dhyiinasamadhi-siltra :!t:flj!..::::.!!*g (T1S, 269c-286a), although

10) Mizuno K6gen 7j(!Jlf5.L5I; distinguishes in his" Introduction to the His­ tory of the Meditation Doctrines in China before the Formation of the Chan School" 1lji*!ilt:s'L£J.MO) /7- 0)1ljitE,I[I,;f~5!:.J'f~3t (Komazawa daigaku kenkyu kiyo ,~rR*"f:1i7fYE{f,C§lf, No. 15, March 1957, p. 20) between a broad sense of the chan jing/zen kyo :fljim£ which encompasses all sources used by the Buddhist practitioner as reference materials for his practice and a narrow sense which refers only to the meditation manuals compiled by the Yogacarins of North­ 'West India. It goes without saying that my usage of the term dhyana sutras belongs to the narrow sense. However, lVlizuno includes many samadhi slltras in his list of chanjing in the narrow sense Cid., pp. 21-23), which seems to me to be a rather loose usage of the term chan jingo Despite all difficulties of distinguishing some chan jing from the sanmei jing, I think it is, nevertheless, more appropriate to treat the two categories separately.

( 44 ) largely based on traditional meditative practices and theories, deals in its last part with the spiritual cultivation of the Bodhisattva and at­ tempts to encompass the two paths in a harmonious pattern. Other texts like Kumarajiva's translations: the Chan mi yaofa jing :flli~§lf~$£ (TIS, 242c-269c), the Siwei lile yaofa }l!kltrua.§lf~ (TIS, 297c-300c); Dharmamitra's ~.~~ (356-442) translation of the Wumen chan jing­ yao yongfa 3i~~:flli$£§lfm~ (TIS, 32sc-333a); Juqu Jingsheng's m.~ Ji'(Ff (5th century AD) translation of the Zhi chanbing mi yaofa ~M}jim ~§lf~ (TIS, 333a-342b); etc., although still indebted more or less to the meditation system proper to Conservative Buddhism, include a considerable number of Mahayana practices and theories or re-inter­ pret traditional methods in a Mahayanist spirit.!l) Although the textual formation of many of these dhyana sfitras re­ mains a very complicated process, the original texts or, at least, much of the meditation practices and doctrines, especially those associated with Conservative Buddhism, can be traced back to the Kashmirian Yogacara school belonging to the Sarvastivada tradition. Most of these meditation manuals were compiled or, at least, reflect the prac­ tice and theory of the Sarvastivada Yogacarins of the first four cen­ turies of our era. We know that long before the rise of the Vijfianavada or Representation-only Yogacara school, a certain group of " masters of spiritual cultivation" or Y ogacarya ~{1Jngijj, specializing in medita­ tion, were active inside the Sarvastivada tradition, especially in Kash­ mir and North-West India.!2) Abhidharma literature, especially the Abhidharma mahavibha$a-sastra, offers an abundance of examples of the Y ogacaryas being quoted as a most reliable source.13) Apart from

11) I have included in the list above only the major extant dhyiina SIUras. 12) On the Kashmirian Yogacara school, v. Paul Demieville, "La Yoga­ carabhilmi de Sailgharak~a," BEFEO, Tome XLIV, Fasc. 2, pp. 340-2. 13) Nishi Giyil W#;-a$, "The Yogacaryas and Their Role in Schismatic Buddhism " ~M19t?ffi:lcfjNt -3 ~1iJp~iji !::: -'t (7):f'kt1j, Bukkyo kenkyi119~?ffi:1i7f~, Vol. 3, No.1 (1939), pp. 1-48.

C 45 ) their role in the formation and development of the Sarvastivada sys­ tem of spiritual practice, the Sarvastivada Yogacarins composed their own meditation manuals and treatises which represent the dhyiina siitras in their original form.1 41 These dhyiina siitras clearly show that the Y ogacarins were more interested in the concrete details of the spiritual training than in the philosophical speculations of the Abhi­ dharma and that, although part of the Sarvastivada tradition, they proved to be open to influnces coming from other schools and Maha­ yana Buddhism. It is quite possible that on the base of their dis­ coveries and insights resulting from their spiritual quest as well as under the pressure of the growing Mahayanist trend around them, this group of " masters of spiritual cultivation" or, at least, a part of them were led to elaborate new theories and practices which, eventually, served as the base of the Representation-only school.1 51 However, much of this historical process of transition from the Sarvastivada Yogacara to the Vijiianavada Yogacara remains to be elucidated. And we must not forget that, as Demieville aptly puts it, "for whoever takes the risk of making the history of Indian Buddhism according to Chinese texts, prudence is a must, not only in chronological mat­ ters."161 As I have already shown, mindfulness of breathing is a central prac­ tice in Conservative Buddhism and it loses much of its importance in Mahayana. A similar phenomenon is to be found in the dhyiina siitras. Aniipiinasmrti is extensively treated in those texts belonging to the orthodox Sarvastivada. These siitras are, therefore, the object of my examination here. On the other hand, mindfulness of breathing tends 14) cf. Nishi Giyu, op. cit., p. 47. 15) Yaoabe Nobuyoshi calls what I term here the" Sarvastivada Yoga­ dira" the" Proto-Yogacara," which denotes the stage before the formation of the Vijfianavada school proper (cf. "An Shigao as Proto-Yogacara," 1991, unpublished manuscript). 16) op. cit., p. 358.

( 46 ) to become a marginal technique in those dhyiina siitras influenced by Mahayana teachings and practices. The Siwei lile yongfa, for instance, makes no mention of the mindfulness of breathing. The Wumen chan jingyao yongfa, although supposed to deal with "the five meditation gates," makes only few brief remarks concerning the iiniipiinasmrti at the beginning of the siitra (T1S, 32Sc), but it gives no further details on it. The Chan mi yaofa jing contains a whole section on the mind­ fulness of breathing (T1S, 2S6c-2S8b), but apart from a few lines dealing with breathing counting, the iinapanasmrti appears to be an auxiliary technique included into the larger frame of the meditation on the impure, which is, however, treated from a Mahayanist viewpoint and has little in common with its counterpart in Conservative Bud­ dhism, as it lays more emphasis on the visualization of different images resulting from this method. Before dealing with the anapanasmrti, we need a few words on the textual history of the main dhyana sz7tras examined here. First, the Da anban shouyi jing represents an extremely corrupt text which gathers together An Shigao's original translation, almost impossible to recon­ struct, fragments from An Shigao's own commentary as well as frag­ ments from glosses by Chen Hui ll*~, Kang Senghui ~{jlff~, Zhi Dun x3li, Daoan 3l!'ti.:, and Xie Fu ~M~. The original text, most prob­ ably written in Gandhari and entitled *Aniipanasvadi, was a small manual of the Sarvastivada Yogacarins compiled around 100 AD. The Parthian monk An Shigao translated this manual into Chinese by the middle of the 2nd century, but the present text dates from the end of the 6th century.17l Sa:qlgharak~a' s Yogacarabhiimi-siitra was trans­ lated into Chinese by Dharmarak~a in 284. Its present text obviously

17) Cf. my article" On the Compilation Process of the Present Text of the Anban shouyi jing Translated by An Shigao " 'ti:'i!!:~~ II''ti:'JiI!t'i'J.Km:EJI ~ff*C1) !1X::lz:lc-::>1,,-C, Toyo no shiso to shukyo J!U:1"(.7),\!.t;mc*~, No.9 (1992), pp. 48- 63.

( 47 ) contains two different siltras.' the first 27 chapters, which represent the Sarvastivada Yogaciira meditation system, and the last three chapters, which are a Mahayanist appendix added in China around the middle of the 4th century most probably in keeping with the growing interest in the teachings of the Greater Vehicle. The Dhyiinasamiidhi-siltra, translated or compiled by Kumarajiva in 407, is a systematical pres­ entation of both Conservative and Mahayanist meditation practices and theories. The chapter dealing with the mindfulness of breathing doubtlessly relies on materials belonging to Sarvastivada. According to Sengrui's preface, which carefully notes the authors whose texts were used in compiling the siltra, "the giithiis dealing with six mental impediments *Jt represent Asvagho~a's practice" (T55, 65b2-3) and " the six aspects ** of [the mindfulness of] breathing gate have been taught by various Vaibha~ikas or Abhidharma masters ~~i!i " (id., b4- 5). We find, indeed, the six mental impediments similarly treated in Asvagho~a's Saundarananda-kiivya, Chapter 15, which actually con­ cludes with the recommendation that one should remove them through the iiniipiinasmrti practice (15; 64). And as we shall see below, the six aspects represent a technique elaborated by Abhidharma masters. Buddhasena's Yogiiciirahhilmi-siltra was composed by the Kashmirian Yogacarya Buddhasena around 400 and translated by his disciple Bud­ dhabhadra sometime after 413. The present text in the TaishO daiziJ kyo is called Damoduola chanjing :i¥.1Hlw.iIi£, as it is wrongly attrib­ uted to Dharmatara or Dharmatrata. It devotes the whole of its first juan to a very detailed discussion of the iiniipiinsmrti, whose practice is divided into four stages, i.e., retrogression m5}, establishment 1:t5}, progression fh1§:5}', and completion i:k:JE:5J"-, each treated from two viewpoints, i.e., method or the preparatory way (prayoga-miirga j]1J!m:) and insight aspects or superior way (vise$amiirga JmJJi!).18)

18) For the textual history of the above texts, cf. Sl!t6 Taishun's {tiii*~ (Continued to next page)

( 48 ) The actual of technique of the mindfulness of breathing described in the dhyana siltras consists of two methods, i.e., ' the sixteen bases' and the ' six aspects.' The sixteen bases represent a very old tech­ nique dating back, most probably, to the earliest days of Buddhism. Lambert Schmithausen and Johannes Bronkhorst point out that mind­ fulness (sati), which in the beginning merely concerned the body, play­ ed an important role in original Buddhism, 'although it may have been borrowed from outside movements, because it appears to be known to Jainism.'19) The subject is too complex and I intend to deal with it on another occassion, but I think we can agree that the sixteen bases of the mindfulness of breathing are a practice peculiar to Bud­ dhism and that they belong to the earliest Buddhist stratum. Cer­ tainly, India has an extremly long tradition of breathing practices, but the Buddhist anapanasmrti must not be confused with a respiration control technique. At no stage does the Buddhist practitioner try to control or change something in the natural process of breathing. Ana­ panasmrti consists of a careful concentration on breathing in and out20)

(Continued from page 48) introductions 1W~ to his translations in the Kokuyaku issai kyo OO/tR-iWi1i£, Kyoshii-bu i1i£~$, vol. IV, 1984 [1928]. and P. Demieville, op. cit., pp. 342-63. 19) Johannes Bronkhorst, The Two Traditions of iVleditation in Ancient India, Franz Steiner Verlag Wiesbaden GmbH, Stuttgart, 1986, p. 89. 20) JInapanasmrti means" mindfulness (smrti) of breathing in (ana) and breathing out (apana)." According to Hattori Masa'aki R~$lEJ3Ij (" Prii1;.ii­ yiima " Wii],i&J" Appendix VIII to the Mahayana Buddhist Scriptures :;k;JI'HLjIJ. (vol. X), Ch116 koronsha 9'l;k;:1}~t±, Tokyo, 1975), ana-apana was originally patterned after the compound praIJa-apana, which in Vedic and classical med­ icine sources stands for "breathing above heart" and "breathing below heart," a distinction made in conformity with the ancient Indian physiology. However, the only allusion to such an etymology I could find in Buddhist literature is the following passage from the Abhidharma Mahavibha$a-sastra (T27, 134a29-b1): "Some people teach that the upper breathing is called iina (' taking in ') and the lower breathing is called apiina (' taking out ') " 1flm

'1f 0 J:.,Ej,1Sf!f*o T,i&J,1Sf!f~o However, the compilers of the iVlahavibhii~ii re­ ject this opinion as incorrect. 'Whatever the etymological protohistory of the ana-apana may have been, I think that since the earliest days Buddhists per­ ceived and interpreted it as " breathing in and out".

( 49 ) with all its physiological and psychological implications. On this ba­ sis, the Buddhist practitioner expands his field of observation to im­ permanence, detachment, etc. Let me quote from the Aniipiinasati szltta (Majjhima Nikiiya 118):

Mindful he breathes in, mindful he breathes out. Whether he breathes in a long (breath) he comprehends (piijiinati) 'I am breathing in a long (breath)'; or whether he is breathing out a long (breath) he comprehends, 'I am breathing out a long (breath) '; or whether he is breathing in a short (breath) he comprehends, ' I am breathing in a short (breath) '; or whether he is breathing out a short (breath) he comprehends, 'I am breathing out a short (breath).' He trains himself (sikkhati), thinking, ' I will breathe in experiencing (patisamvedi) the whole body.' He trains himself, thinking, ' I will breathe out experi­ encing the whole body.' He trains himself, thinking, 'I will breathe in tranquillising (passambhayam) the activity of the body (kiiyasamkhiira).' He trains himself, thinking, ' I will breathe out tranquillising the activity of the body.' He trains himself, think­ ing, ' I will breathe in ... breathe out experiencing rapture (piti).' He trains himself, thinking, 'I will breathe in ... breathe out experiencing joy (sukha).' He trains himself, thinking, ' I will breathe in ... breathe out experiencing the activity of thought (cittasamkhiira) ... tranquillising the activity of thought ... ex­ periencing thought (citta) ... rejoicing (abhippamodayam) in thought ... concentrating (samiidaham) thought ... freeing (vi­ mocayam) thought.' He trains himself, thinking. 'I will breathe in ... breathe out beholding (anupassi) impermanence (anicca) ... beholding detachment (viriiga) ... beholding stopping (ni-

( 50 ) rodha) ... beholding casting away (patinissagga).'21)

These sixteen stages or sixteen bases (solasavatthuka), as Buddha­ ghosa calls them,22) are found in all the dhyana siitras that discuss the anapanasmrti. They are called " the sixteen excellent [practices] " + ~1lJ in An Shigao's translation of the *Anapanasvadi (=As) (cf. T1S, 16Sa8-19) and +~*f1lJ in Sarp.gharak~a's Yogacarabhiimi translated by Dharmarak~a (=SYb) (cf. T1S, 216a1S-28)23) or" the sixteen limbs" +~7J'- (*~o¢asanga) in the Dhyanasamadhi-siitra (=Ds) (cf. T1S, 27Sb19-276aS) and Buddhasena's Yogacarabhiimi (=BYb) (cf. T1S, 31Ob22). However, the same BYb also uses the translation +~ff (*§o¢asacara) (cf. T1S, 302b1-14). On the whole, these sixteen prac­ tices are similar to the original pattern found in the P1ili text above. A detailed discussion of each particular text is not possible here. In what follows I shall point out the most conspicuous differences. The As seems to contain the largest number of differences, but the text is so unclear and corrupt that it is difficult to decide whether it repre­ sents a variant or just a very clumsy rendering into Chinese. The

21) I. B. Horner's translation in The Collection of the Middle Length Say­ ings, Vol. III, Pali Text Society (=PTS), Luzac & Company, LTD., London, 1967, pp. 124-125. (I have introduced the Piili terms in brackets, which do not exist in Horner's translation.) The original text is in Majjhima-Nikiiya, ed. by Robert Chalmers, Vol. III, PTS, Luzac & Company, LTD., London, 1960, pp. 82-83. The whole sutta as such consists of two historicalIy different layers and was compiled in the present form one or two centuries after Asoka's reign. The part which contains the passage above is very old or at least trans­ mits a very old practice dating back to the earliest days of Buddhism. A simi­ lar text, repeated in different versions, is found in the Sa1?zyutta Nikiiya 54, Aniipiina Sa1?Zyuttam (cf. Sa1?Zyutta-Nikiiya, ed. by Leon Feer, Luzac & Com­ pany LTD., London, 1960, pp. 311-341). 22) Cf. The Vissudhi-Jl!lagga, ed. by C.A.F. Rhys Davids, PTS, London, 1975, p. 267, 291. 23) Demieville, op. cit., p. 415, n. (1), attempts to reconstruct iflfJl91 as *uttaracaryii making reference to Xuanzang's 34~ translation as Mfr.

( 51 ) order of the description is not identical to the one found in the Piili text: after the stage of comprehending the movement of the body JJJ $1', the practitioner returns to respiration and observes whether he breathes faintly 1*;, whether his breath has stopped or not 11:: • /F11::, etc. The text says nothing about experiencing, etc., the activity of thought. The impermanence is not translated by the usual ~~, but by a periphrasis: " he breathes and comprehends that all the things in this world (' the ten thousand things ') [are bound to] perish and there is no way to obtain them back" ;ft7J4We$/FRJ:fj[1~Piffi,~,r3~. The text ends with the stage of beholding detachment ~mffl,~' and stop­ ping ;Qfl:~~ilJ, which means that it lacks the patinissagga phase of the Piili text. We also find some differences in the Ds, which lacks some of the stages pertaining to the activity of the thought and the thought described in the Aniipiinasati sutta. Besides, the thirteenth stage, which is called" the observation of vanishing l±l1*1lJll, of the conditioned (sa1!lskrta) dharmas 1f~*," has no equivalent in the Piili source. The above dhyiina siitras as well as many of the Abhidharma treatises also contain a second technique of the iiniipiinasmrti called "the six aspects" or " the six means." I shall first present" the six means" (~arjkiirafla 1\ r:sJ) according to the Abhidharmakosa-bhii~ya I\PJ mJiL'i~H~ ~jjffi;,24) since Vasubandhu's systematic way of explaining them will allow us to understand the variants appearing in the dhyiina siitras and Abhidharma sources. The practice starts with "counting" (gmJanii ~), which consists in counting breathing from one to ten. When this is accomplished without any counting failure (do~ii !J2..), the practioner advances to the second step, i.e., "pursuing" (anugama ~), which means intently following the inhalation as it enters the body and moves from the throat, through the heart, the navel, the kydneys, the thighs to the toes and then the reverse movement of the exhalation until it

24) Abhidharma-koshabhiifya of Vasubandhu, ed. P. Pradhan, Patna, 1967, p. 339-340; Xuanzang's Chinese translation is found in T29, 118a23-b15.

( 52 J leaves the body. Next comes" concentration" (sthiipanii ll) which denotes focusing one's attention on some part of the body from the tip of the nose to the big toe. In the fourth step, called" observation" (upalak$anii jjt), the practitioner discerns that the air breathed in and out as well as form (rupa {5), mind (citta Jt,), and mental functions (caitta Jt,pjf) ultimately consists of the four great [basic] elements (mahiibhuta ::kfll!). He thus analyzes all the five aggregates (paiica skandhii/; JL.fi:). Next follows" the turning away" (vivarta ~) which consists of changing the object of observation from the air breathed in and out to "the wholesome roots" of purity (kusalamula ~fR) and ultimately to " the highest mundane dharma" (aggradharma tl!:Flmrf~­ ~). The last step is called "purification" (parisuddhi ~) and it marks entering the stage of" realization of the Way" (darsanamiirga JtJ!t), which in Abhidharma literature denotes the stage of" the stream entry" (srota iipatti-phalla) that will inevitably lead the adept to

NirviilJa in no more than seven lives. 25) Before examining this technique in the dhyiina sutras, it is necessary to note that the sixteen bases and the six means represent historically different practices. It is quite obvious that both methods lead the practitioner to very high stages of spiritual achievement. It is very hard to imagine that a practitioner who has reached pa/inissagga or darsanamiirga needs to return to such elementary techniques like counting or mere mindfulness of the length of breathing respectively. Judging from a logical viewpoint, we are bound to say that the two techniques are independent and seem to lead to approximatively equiv-

25) The six means are also treated in the Abhidharma mahiivibhii~ii-siistra JliiJm:i¥Jll!*~rJ>~ (T27, 134c26-135b17), the SG1'[lyukta abhidharmahrdaya­ siistra ~JliiJm.,L'~ (T28, 934a23-bl0), the Abhidharma nyiiyiinusiira-siistra JliiJm:i¥Jll!JiJjtIE~ (T29, 673c10-674a24). In the Theraviida tradition, the Vi­ suddhi-magga divides the same practice into eight stages called" the mental [training] methods" (manasikiiravidhi) (cf. op. cit., pp. 278-87). For others sources, see below.

( 53 ) alent stages of spiritual achievment. The sixteen bases go back to the earliest stratum of Buddhism, while the six means date not earlier than the 2nd or 1st century Be. The Pali suttas contain no reference to anything similar to the six means26) or ga1)ana in the sense of breathing counting.27) The seven fundamental Sarvastivada Abhidharma trea­ tises contain no mention of the six means, but we must bear in mind that breathing and anapanasmrti do not get an extensive treatment here. It is, therefore, unclear whether the six means did not exist at the time of the compilation of the oldest of these texts or whether they were still a minor practice unworthy of the attention of the Sarvastivadin masters. The earliest mention or, at least, the record of what appears to be the earliest form of this practice is found in the Vimutti-magga M&Ji~ (T32, 430bI7-29). The dating of this text is very difficult, yet the 1st century AD seems to be the most probable date.28) Its author Upatissa refers to "the four ways of practicing the mindfulness of breathing" IZ9fl{Iil:~:R%t as the teaching of " the ancient masters" (pubbacariya %~!ji), which suggests that the technique was older than the date of the composition of this treatise. Furthermore, we have here only four methods of practice, i.e. 'fIf., Ilifj~, :R@:, and llifjiliJL which correspond to the first four stages of the six means. Not only that we do not have the last two steps, but" observation" Ilifjll. refers here to the mindfulness of such psychological states as joy :g, rapture ~, etc., and not of the four great elements. All these facts prove that we have here a very early model of the six means, which I would ven-

26) Cf. Tamaki Koshiro .:E~Jl!tJmE!~, " Fundamental Problems concerning the JIniipiina-samiidhi" )\1±\)[!l,~tE0)1l*Fp~M:l in Collected Essays of Buddhist Studies Presented in Honor of Professors Ito Shinjo and Tanaka Junsho 1jJtJii~ :9tt • EEjI:pJIIi'iJffijjllj~d:¥:~t\!j~~{'M!t'j€:ffilii)(~, Toho Shuppan, 1979, p. 80. 27) lowe this information to Ven. Sumanasara Thera who also suggests the possibility that breathing counting was borrowed from: Hinduist practices. 28) Cf. Nakamura Hajime, Indian Buddhism: A SZirvey with Bibliograph­ ical Notes, Motilal Banarsidass Publishers, Delhi, 1987, p. 116.

( 54 ) ture to place somewhere in the 1st century BC. Since the six means/ eight methods practice is shared by both the Theravada and Sarva­ stivada traditions, it is possible that the primitive pattern dates from a period prior to their schism. Of course, there is also the possibility of its being created by one sect and borrowed by the other, but this must have happened at very early date, too. However, if we accept the first alternative, then the origin of this practice may be placed as early as the 2nd century BC. There is no doubt, however, that this practice was originally an auxiliary exercise devised by the Abhidharma masters. The second stage in the historical development of the six means practice is reflected in the As and the SYb. In both siitras, we find that the practice has now all the six stages, but the newly added steps are, however, explained in a manner different from the Kosa. The As text describing the last two stages of" the six aspects" A-* (*$a4vastu) is extremely corrupt and our interpretation is in danger of taking a later Chinese gloss for the original siitra, but it seems that " turning away" ~ corresponds to the stage of "practicing the Way" 1-TJH (TIS, 16Sa28) and" purification" l$- represents" entering the Way" AJH (TIS, 16Sa28-bl). ~ and l$- are also explained as " eliminating the fetters (sa1flyojana) " *~ (TIS, 167aI9), which, if we are to be­ lieve the Chinese gloss, itself ambiguous and contradictory, would mean casting away of the physical and verbal defilements and elimina­ tion of the mental defilements respectively. Furthermore, unlike the Vimutti-magga, "observation" if)! is now explained as observing the five aggregates (TIS, 167a7-8). The SYb speaks about "four as­ pects" 12:9*, but in realitY"we have six stages, i.e. )~,~, ;t~IliR, ll::, ~, ~, l$- (TIS, 216a29-b23), counted in a different way. Their ex­ planation is very summary and rather unclear. The last two stages are described as "fixing one's attention on the tip of the nose, one must observe the breathing counting and be aware of the breathing out

( 55 ) and in" (T1S, 216b18-19), which represents an explanation much earlier than the Kosa. Both the As and the SYb reflect the six aspects practice around 100 AD, when the last two stages had not been fully co-related with Abhidharma categories. The last phase starts with the Mahiivibhii~ii, from the middle of the 2nd century AD on, when all the stages of the six means practice, despite certain differences between various texts, become very well defined and included within the larger frame of the Abhidharma the­ ories of the spiritual path. Most of the differences concern the inter­ pretation of the last two stages. The Ds and the BYs as well as most of the Abhidharma sources belong to this third phase of development of the six means. The Ds calls this practice "the six gates of the iiniipiinasamiidhi " 1\PJj3~itl-5tj3~:=:,!!;1d~f.i F~ (* iiniipiinasamiidhi ~acJ-naya) (TIS, 273aI8-I9) and describes the fifth stage, i.e., if!iiiJL as observa­ tion of the impermanence of the five aggregates which leads to the elimination of the five obstacles 3i!l: (pafica-nivarat}iini). Purification is described in relation to the practice of the four fields of mindfulness ug~Jl (catviiri smrty-upasthiiniini),29l which leads to the attainment of the four wholesome roots ug~1R and finally to the Arhatship ~¥: (TIS, 27Sb7-19). The BYs, which gives the most detailed treatment of the six means, holds that" interior attachment " 1*Jji:~ disappears when counting is perfectly practiced (TIS, 307c8), that" exterior at­ tachment" 7i-ji:~ is eliminated at the stage of pursuing Il1H (TIS, 306blO), and that doubt ~ is cut off at the stage of observation ~ (TIS, 307a29). The turning away ~if!i is presented as the stage of cultivating wisdom ~~ (prajfiii) (TIS, 307bI6). The BYs considers that a purification is obtained after each of the previous levels, but purification as a stage in itself is described as the cessation of all the evils JllJ~ which have constituted the base fiJf{l\;: of the defiled life

29) In the Mahavibhii1ii (T27, 12Sa23-2S), it is at the stage of turning away ijii that the adept practices the four bases of mindfulness.

( 56 ) (cf. T1S, 307cS-17). How did the Y ogadirins correlate the sixteen bases with the six means, which originally had been independent, parallel techniques, in actual practice? The As, the SYb, and the BYs as well as the Vimutti­ magga, the Mahiivibhii$ii, and the Abhidharma nyiiyiinusiira-siistra sim­ ply describe the two practices without giving any detail on how they relate to each other. The Ds, in a passage rather difficult to interpret (T1S, 27Sb19-20), seems to suggest that the six means are included in the first step of breathing in and out of the sixteen bases. In a similar way, the Visuddhi-magga tries to include the eight mental methods into the sixteen bases, between the first and the second tetrads. Yet we must never forget that many details concerning the actual practice were transmitted orally and are impossible to reconstruct now. What we can say in conclusion is that Buddhism, as many other religions, has often found creative solutions by trying to harmonize what originally represented different or even contradictory practices.

( 57 )